18, 2 King: Philippine Concrete and its Aggregates 207 
should be used with caution. Mariquina River sand is much 
coarser than that from Pasig River and yields strong mortar 
of appreciably higher strengths than those given by standard 
Ottawa sand, as Table 2 shows. Sand from Mariquina River 
contains practically no quartz and has been derived from the 
weathering and erosion of andesitic and basaltic rocks. The 
grains are fairly soft, but excellent results have been obtained 
from laboratory-made concretes containing this sand, and if due 
care be exercised there is no reason why Mariquina sand should 
not give very good results in the field. 
Concrete tests, made in connection with the fortification of 
El Fraile and Carabao Islands at the mouth of Manila Bay, 
show that laboratory-made specimens are much stronger than 
those coming from Fort Mills, and those made of 1: 254 
concrete at an age of 28 to 36 days have a mean compressive 
strength of 2,693 pounds per square inch. Older field specimens 
made of 1: 2:4 concrete, and aged 28 to 38 days, have a mean 
compressive strength of 1,104 pounds per square inch. Still 
older cubes made in the field show less strength than this; those 
44 to 48 days old average only 907 pounds per square inch. A 
comparison of the compressive strengths of concrete specimens 
coming from Fort Mills with those of the provincial test pieces 
that gave the best, results shows the marked superiority of the 
latter. With the materials available, much better results should 
have been obtained from the field mixtures here recorded. 
MARINDUQUE 
With one exception, all results obtained from concrete speci- 
mens coming from Marinduque are very poor. Six cubes made 
on August 19, 1915, of 1 : 2 : 4 concrete used in constructing 
Tiguion Bridge, at Gasan, are the only specimens tested. The 
average value shown is 443 pounds per square inch; the min- 
imum, 181; and the maximum, 1,014—results that are certainly 
extraordinarily erratic. Examination of the spalls and frag- 
ments remaining after testing these cubes showed not only that 
a fine beach sand was used, but also that there was unmistakable 
deficiency of cement, so that in reality the concrete contained 
less cement than that required for a 1:2:4 mixture. Insuf- 
ficiency of cement is therefore the chief cause for this very low 
average value of 443 pounds per square inch, which is just 
about one-fourth of what a good 1 : 2 : 4 concrete should test at 
28 days. Varying percentages of cement probably account for 
the freakish results, no two of which are alike, though nominally 
